LESSONS FROM THE US INCENTIVE AUCTION

LESSONS FROM THE US INCENTIVE AUCTION The dust has settled on the world s first auction that incentivised broadcasters to relinquish spectrum for wireless use. HOWARD SYMONS and PAUL MILGROM, who were part of the auction team, reflect on the process and the implications Consumer demand for wireless communications has exploded, and mobile data usage continues to grow exponentially. Nations and carriers face a major challenge to ensure that the speed, capacity and accessibility of wireless networks keep pace with these demands in the years ahead. Meeting this challenge requires creative solutions for making more spectrum available for mobile broadband through the reallocation of frequencies, spectrum sharing and the authorisation of new secondary uses on bands that are otherwise underutilised for their primary purpose. One such solution is the use of the two-sided incentive auction, which uses market mechanisms to identify supply as well as demand. The US successfully used such an auction to repurpose 84 MHz of spectrum from television broadcasting to mobile broadband, which concluded in April 2017. In addition to reallocating a substantial amount of spectrum, the auction produced a surplus of nearly $7 billion, which will be contributed to the US Treasury. This article discusses the US incentive auction, the challenges related to participation and auction design, and the relevance and applicability of two-sided spectrum auctions outside the US. PURPOSES AND GOALS OF THE INCENTIVE AUCTION A spectrum incentive auction is a market-based means of repurposing spectrum by encouraging licensees to voluntarily relinquish spectrum usage rights in exchange for a share of the proceeds from an auction of new licences to use the repurposed spectrum. The incentive auction combines a reverse auction, in which incumbents bid to relinquish their spectrum usage rights; a forward auction, in which buyers bid to acquire the relinquished spectrum for new purposes; and a procedure to reconcile the two. Auctions have been used by the US Federal Communications Commission (FCC) since 1994 to assign new licences. Before the incentive auction, the FCC simply used its plenary power to relocate the incumbent users from spectrum bands that were licensed via auction. The incumbents were not given any financial incentives to move (other December/January 2018 Vol 45 Issue 4 InterMEDIA 25

Broadcasters Reverse auction Before After REVERSE AND FORWARD Broadcasters relinquish spectrum usage rights for compensation Mobile broadband providers bid for spectrum licences THE REALLOCATION Mobile broadband providers Forward auction than access to a mechanism for reimbursing them for the costs of moving). The demand by mobile providers for more low-band spectrum, which can transmit longer distances and penetrate walls for indoor coverage, called for a different approach. In the US, broadcasters do not own their spectrum; they use it under licence from the federal government. Nonetheless, the FCC faced political resistance to moving the broadcasters from their channels in the 600 MHz band. The answer was to give broadcasters a financial stake in reallocating the broadcast band by passing a law giving broadcasters the right to seek compensation through the auction for relinquishing their frequencies. The auction, which formally closed on 13 April 2017, was a success. TV channels 38-51 (614 698 MHz) were reallocated, with 70 MHz for use by wireless carriers and 14 MHz for unlicensed uses. This reallocation was accomplished by purchasing the spectrum of 175 broadcasters in the reverse action for a total of $10 billion, paid for out of the proceeds from the sale of this spectrum to wireless carriers in the forward auction. The 50 winning bidders in the forward auction paid more than $19 billion for licences in the repurposed 600 MHz band. In addition to funding the purchase of spectrum usage rights from the broadcasters, the forward auction receipts covered the establishment of a fund to compensate the costs of moving 957 of the broadcasters remaining on the air post-auction to new channels, as well as the FCC s costs in developing and running the auction. THE CHALLENGE OF STAKEHOLDER PARTICIPATION While ultimately successful, the auction faced several challenges. One significant challenge was that broadcasters were sceptical of the auction and there was a risk that insufficient numbers of them would participate. The US has about 2,000 television broadcast stations. While many are owned by large companies, licensees include small and family-owned businesses. Broadcasters, even the larger group owners, were unfamiliar with spectrum auctions. Many were concerned that the auction would not be truly voluntary, while non-participants in particular feared losing coverage after the auction. Some broadcasters thought they could reallocate their frequencies for mobile use through private transactions and the introduction of the new ATSC 3.0 advanced television transmission standard, which will allow for mobile transmissions on broadcast spectrum. Broadcasters therefore needed to be persuaded that national reallocation by the FCC (the sole organisation with the authority to reorganise the band on a nationally uniform basis) was the only effective means for creating the necessary scale and scope for new 600 MHz ecosystem, which in turn would maximise the value of the broadcast spectrum tendered in the reverse auction, making each broadcaster a valuable part of a larger whole. Reallocation required the coordinated development of standards so users would not interfere with each other at the boundaries of their licence regions; international coordination to avoid treaty issues with Canada and Mexico; and the establishment of guard bands to prevent interference with users in adjacent frequency bands none of which could have been accomplished on a nationwide basis by broadcasters acting through the secondary market. To encourage broadcaster participation, the statute authorising the incentive auction gave them three bidding options: a participating broadcaster could choose to bid to relinquish its spectrum usage rights and go off the air; retain its licence and broadcast from a shared channel; or move to a lower frequency band (from UHF to VHF, for instance, or from a high VHF channel to a low VHF channel). The statute also gave the FCC authority to reassign ( repack ) TV stations that remain on the air so that they would occupy a smaller portion of the UHF band. But the repack was designed to allay concerns by broadcasters that the reassignment to a new channel would adversely affect their business. Non-participating broadcast stations and broadcasters whose bids were unsuccessful were guaranteed a channel in their pre-auction band. Congress also required the FCC to take all reasonable efforts to preserve the coverage area and population served of stations that would remain on the air after the auction, with those metrics set as of 22 February 2012, the date of enactment. This Former interference area REPACKING RULES B-Ch.21 Former interference contour C-Ch.50> C-Ch.21 New interference contour A-Ch.20 Noise limited contour Unique new interference area not permitted unless <0.5% A schematic of the rules governing the repacking of broadcasters onto new channels after the auction. Assignment to a new channel is not permitted if the population served loss exceeds 0.5% (yellow); the predicted areas of no change are indicated in black. The new coverage area will replicate a station s contour on its new channel. 26 InterMEDIA December/January 2018 Vol 45 Issue 4

authority enabled the FCC to reconfigure a portion of the UHF band into contiguous blocks of spectrum suitable for wireless use, while assuring broadcasters that there was no risk to participating in the auction. All of these constraints also created challenges for auction design, described in more detail below. Finally, Congress also established a $1.75 billion TV Broadcaster Relocation Fund to pay the reasonable relocation costs of television stations that moved to a new channel in the smaller post-auction 600 MHz band allocated for broadcasting. FCC staff developed a list of eligible expenses and estimated costs with a goal to balance expediency with the avoidance of waste, fraud and abuse. To educate broadcasters and encourage them to consider participating in the auction, the FCC mounted an unprecedented outreach campaign. Educating broadcasters about the incentive auction is analogous to providing potential investors with the information they need to make a knowledgeable decision about whether to fund a multibillion dollar financial transaction. The FCC engaged an investment banker to assist in this effort, and prepared an information package for every auctioneligible station owner. This package included high-end estimated compensation values for every television market. FCC staff visited 30 cities around the US and met with hundreds of broadcasters to explain the auction and address questions. The FCC also adopted a number of policies intended to encourage forward auction participation by a wide array of bidders to promote more competition in the marketplace for mobility services. Foremost among these policies was a set-aside of up to 30 MHz of spectrum in the forward auction for bidders other than the largest two carriers. This conditional reserve is described in more detail below. Additional measures included small business bidding credits and credit for small carriers serving rural areas. AUCTION DESIGN CHALLENGES There were many new challenges in designing the rules of the US incentive auction, among which three stand out as particularly difficult, requiring novel solutions. The first of these challenges was to keep the auction relatively simple for the broadcasters, so that they would trust the process and participate. What made this hard was that the auction outcome needed to respect interference constraints, which limited the way channels could be assigned to stations. There were literally millions of such constraints, mostly of the form one cannot assign station X to channel A and station Y to channel B. Deciding whether there is some way to assign channels so that any particular combination of stations could all continue to broadcast is very hard computationally so hard that bidders would likely be unable to check that the auctioneer s computations are correct. In such a setting, a small broadcaster cannot know how high a bid must be to win, so how can it think about how to bid? Is there any way that the inherent complexity of the auction s machinery can be kept under the hood and out of the view of the bidders? To ease the broadcasters bidding problem, the FCC adopted a simple auction rule: a descending clock auction. To describe this auction simply, consider the case of a small broadcaster that has decided not to switch to a VHF frequency, but instead decided to either sell its broadcast rights if that can be done at a price higher than X or to keep its rights if it cannot get at least that price. In the descending clock auction, the station s price starts high and declines from round to round. It falls whenever there are enough other willing sellers such that there is still room in the band to accommodate the station. That determination is very hard computationally, but the station s bidding problem is easy. It should continue to offer its rights for sale as long as the currently offered price is any price >X, and it should exit the auction if price is <X. It is obvious that this strategy sells the rights when the station can meet its price target, and not in any other case. The second novel challenge concerned how to clear the market. In a competitive auction market for an ordinary commodity, there is just one price. The same thing is being bought and sold, so what the seller receives is what the buyer pays, minus any commissions. In the incentive auction, however, the two sides are buying and selling very different things: the stations sell their spectrum rights, which are then used to clear some channels in the UHF band. Some of the cleared channels are used to create mobile broadband licences, which are to be sold, but some go to unlicensed uses, such as wireless microphones. The challenge is how to ensure that the revenue from the forward auction is sufficient to pay not only In the incentive auction, the two sides are buying and selling very different things. the cost of the reverse auction, but also the costs of running the auction and moving the continuing broadcasters to new channels, as well as meeting the FCC s minimum reserve price. To accomplish all these things, the FCC adopted the following procedures. The first step took place long in advance of the auction, when the FCC determined how many 10 MHz broadband licences it could sell for any number of broadcast channels to be cleared. It planned to sell 10 broadband licences if 21 channels could be cleared, 9 broadband licences if 19 channels could be cleared, and so on. The detailed plan is depicted on p28. Next, the FCC announced high opening prices for stations, and required stations to declare whether they wished to bid in the auction and to commit to sell at the high opening prices. Then, the FCC computed the largest number of channels that it could possibly clear if all these stations were bought, which established the initial (channel) clearing target. Using the clearing target and the associated number of broadband licences, the reverse and forward auctions would be run. December/January 2018 Vol 45 Issue 4 InterMEDIA 27

BAND PLANS If the net revenue was not sufficiently high or the average price in the forward auction failed to meet a minimum threshold, then the clearing target would be reduced and the auction would continue. If those requirements were satisfied, the final stage rule would be triggered, and the number of channels to be cleared would thereafter remain unchanged. The third novel challenge was how to reserve some spectrum for bidders, other than the largest incumbents, which did not yet have sufficient low-frequency spectrum to provide effective wide-area coverage. Without knowing how much spectrum would be cleared in the auction, how would the regulator know how much to reserve? If it set aside too much and kept the prices low, how could it avoid causing the whole incentive auction to fail, with too little revenue in the forward auction to pay the prices demanded by broadcasters in the reverse auction? The solution to the third challenge was the FCC s conditional reserve. Until the final stage rule, described above, was triggered, there would be no reserve: no spectrum set aside for qualified bidders. All bidders would have to contribute to paying the cost of clearing the broadcasters from the spectrum. However, once the final stage rule was triggered, there would be reserved spectrum created in each licence area. The number of reserved 10 MHz broadband licences in each area would be the smaller of three licences or the number of such licences with bids from qualified bidders. This rule ensured that: l Incumbents would be unable to shut the other bidders out of the bidding l As much as 30 MHz in each area would be provided to non-incumbents l Forward auction prices would be high enough to pay the broadcasters for their spectrum rights, sold in the reverse auction. REGULATORY DESIGN In parallel with educating bidders and developing the auction design, the FCC had to translate the incentive auction into regulation. One critical policy decision was the band plan for the reallocated spectrum. The FCC adopted what it The band plan scenario associated with each potential spectrum clearing target, along with the number of paired blocks offered in each band plan called a down from [channel] 51 plan, with all paired 5x5 megahertz blocks. Blocks above the 11 MHz duplex gap were reserved for uplink transmissions; blocks below the duplex gap would be used as downlinks. Because the FCC would not know how much spectrum would be available to reallocate until the auction began, it adopted band plans for a range of possible clearing targets (shown above). International coordination along the borders with Mexico and Canada was also critical to the success of the incentive auction. Mexico adopted a plan to relocate all of its television stations below channel 37, and Mexican and US regulators confirmed through an exchange of technical coordination letters that the two nations intended to follow common guidelines for repacking TV stations. Industry Canada adopted the same 600 MHz band plan as the US and reached an agreement with the FCC to jointly repack TV stations on the same timetable. Canada and the US agreed to a statement of intent setting out a framework and timeline for repurposing TV spectrum for mobile broadband on both sides of the border. An important element of regulatory design was the post-auction transition. The FCC established a transition period of 39 months for reassigned stations to cease operation on pre-auction channels, with individual deadlines for reassigned stations to move to their new channels in the smaller portion of the 600 MHz band allocated to broadcasting. The FCC has developed a schedule for this repacking that takes account of the complex interference relationships among television stations in adjacent markets (the daisy-chain ). Finally, implementation issues also included reallocation of other users in the 600 MHz band, including wireless microphones and low power television licensees. The FCC s transition rules allow low power stations to operate in the newly licensed wireless spectrum unless and until displaced, while wireless microphones and unlicensed devices may International coordination along the borders with Mexico and Canada was also critical. 28 InterMEDIA December/January 2018 Vol 45 Issue 4

operate in cleared spectrum during the transition period. In addition, the FCC determined that a proportion of the duplex gap would be available for unlicensed devices. APPLICABILITY AND RELEVANCE OUTSIDE THE US While certain aspects of the US solution were subject to criticism, the US experience offers some important lessons to legislators, regulators and stakeholders. The US incentive auction facilitated the reallocation of spectrum by giving sticky incumbents in this case, broadcasters a financial stake in the repurposing of a band. But the incentive auction concept can be used in any band, not just broadcasting; and it can be used to facilitate spectrum sharing and alternative uses of underutilised bands, as well as reallocation. In all of these instances, an incentive auction can promote the efficient use of spectrum without relying on government fiat. Instead, it uses market forces to enable incumbents and prospective users to determine on a voluntary basis whether there are higher value uses for a band. It also creates a win-win-win process, in which the old and new users and the public treasury can all benefit while avoiding, or at least minimising, the political friction that can occur when these determinations are imposed in a top-down fashion. At the same time, policymakers retain the important role in identifying bands that may be the most appropriate targets for applying the auction mechanism to and addressing the questions of reallocation or sharing. Some have noted that the US incentive auction was complex and time-consuming (applications for the reverse auction opened in January 2016 and the auction closed in April 2017). But many of the complex elements of the US incentive auction were the result of policy decisions that need not be part of every such auction. These included making participation by broadcasters voluntary; protecting the service areas and population of broadcasters remaining on the air; providing participating broadcasters with a range of bidding options; and establishing a duplex gap for unlicensed use that consisted of spectrum that had to be purchased from broadcasters but, by definition, was not available to be sold in the forward auction. Nonetheless, the US experience does suggest several elements required for a successful auction. Most fundamental to success, of course, are a sound auction design and a workable regulatory scheme. But of nearly equal importance is creating the environment in which an incentive auction is regarded as an appropriate and useful solution to the vexing problem of continued spectrum scarcity. Creating that environment requires educating policymakers and stakeholders. Policymakers are familiar with spectrum auctions, but applying market forces to identify supply as well as demand will be new, as it was in the US. The US experience teaches that legislators and regulators will want to be assured that an incentive auction can be a workable solution that benefits the public by efficiently and effectively meeting the demand for mobile broadband. The SPECTRUM OF VIEWS ON THE AUCTION Since the auction s end in April 2017, reception has been mixed on the extent of its success and how it could apply elsewhere. l There has been much respect for the FCC for conducting such a complex auction, and recognition of the political realities, but some have questioned the administrative costs and results. They note that the auction did not meet the highest clearing target (70 MHz out of 120 MHz targeted), and the unlicensed use, say for white spaces, was there anyway. Conversely, others argue that the approach reduced the availability of over-the-air television service. Campaigners for public accountability have also complained that winning bidders obtained payment for what should have been public-owned guard spectrum (although this is true for any use of the incentive auction mechanism), and those who elected for channel sharing could also retain their licences and service rights. Broadcasters will also benefit from the transition to ATSC 3.0 technology and from subsidies (see, for example bit.ly/2jbuql8 and bit.ly/2khhxur). l On international applicability, Analysys Mason notes that US broadcast spectrum is much more in the hands of commercial broadcasters than in most other countries, where public service broadcasters hold most spectrum, although there are multiplex operators that serve commercial outfits. They also note that some US operators opted to relocate to the VHF band, which was abandoned long ago in most countries, and that in the US, terrestrial TV is in any case a fairly small part of the market (16% on primary TV sets) compared with cable and satellite, so a loss of audience from stations going off-air is much less than, say, in Italy or Spain. But there are countries such as Germany and Japan that have an even smaller proportion of terrestrial TV households than the US. See bit.ly/2bfzamm. However, applying the incentive auction to non-broadcast bands would not necessarily raise these issues. l Meanwhile the National Association of Broadcasters in the US has launched TVAnswers.org, a website to inform Americans about how they can continue receiving their local broadcast TV stations during the repack process. l An interesting paper (co-authored by Paul Milgrom) shows how economics and computer modelling were combined to work on the incentive auction. Titled Economics and computer science of a radio spectrum reallocation it discusses the essential and unusual challenge of how to make the auction simple enough for human participants while still ensuring that the computations would be tractable and capable of delivering nearly efficient outcomes. See bit.ly/2kgu9ow l Finally, Thomas Hazlett, whose book The Political Spectrum we reviewed recently, and a free market thinker, lost no time in mentioning his preferred market mechanism to free up more spectrum overlay rights. Overlays do not require congressional authorisation, complex auction designs, or contentious uprooting of incumbents, he says. See brook.gs/2kvaggi Marc Beishon details of that experience also suggest the sorts of statutes and rules that are needed to allow the auction to achieve those objectives. Most stakeholders will also be unfamiliar with a two-sided auction for spectrum rights. To encourage stakeholders to participate in such an auction, they need to see it as a fair mechanism. There is no single model for a successful incentive auction. It can be tailored for use on a national, regional or continental basis, respecting differing national uses for particular bands. The US experience was a proof of concept under challenging circumstances. The incentive auction should be a tool in every spectrum manager s toolbox. HOWARD SYMONS is a partner at Jenner & Block. He was general counsel of the FCC and vice chair of the FCC s Incentive Auction Task Force. PAUL MILGROM is the Shirley and Leonard Ely professor of humanities and sciences at Stanford University, and chair of Auctionomics, which designed and helped implement the auction. December/January 2018 Vol 45 Issue 4 InterMEDIA 29